I used a computer processor cooling heat exchanger, computer fan, and a 3.8L/min 12V water pump.  Latex tubing from the pump to and from the heat exchanger and an insulated box. Use either warm water or ice water as conditions dictate.

 

My reasoning is twofold.  In the summer the mirror may be heat soaked from the heat of the day and cause tube currents because the air temperature drops faster than the mirror temperature.  The mirror can take several hours to cool, and always lags warmer than the air until the air temperature stabilizes and the mirror catches up, which might not happen 'till the middle of the night.  Pumping ice water through the heat exchanger allows cooling the mirror quickly down to a temperature at which there are no tube currents.  Now it is probably true that the mirror will see some stresses as a result which will affect it's figure, but I'm also quite sure the view is better than with huge boiling air coming off the mirror.  Without this technique, looking at Jupiter for example is an indistinct fuzzy blob early in the evening.  Cooling the mirror quickly allows the beautiful details of festoons to be observed far sooner than waiting for the mirror to cool by blowing ambient air on the back of it.

 

The other case is in the winter, when the mirror may cool below the dew point.  Now it is true that a Newtonian telescope with the mirror at the bottom of a long tube is less prone to the mirror condensing moisture out of the air than an SCT's corrector plate right out there in the open, but it still can condense moisture.  Not dripping wet, but enough to degrade the view.  In this case warm water can be pumped through the heat exchanger to warm the mirror just enough to keep it above the dew point. 

 

For early cooling during hot weather, I just throw some ice in the water.  For warming the water in the winter, I use a 12V coffee cup heater probe.